How to Build a Grocery Delivery App in 2026: Catalog, Picking and Slots

Why grocery is harder than restaurant delivery

If you have read our guide on how to build a food delivery app, much of the courier side will look familiar. The store side does not. Five things change:

• The catalog is huge. A restaurant menu has dozens of items. A supermarket has tens of thousands, with sizes, brands, allergens and images that must be right.
• Prices are not final at checkout. Bananas, deli meat and steaks are sold by weight. The real price is known only after picking.
• Stock counts are often wrong. Shelf stock in a store drifts from what the POS says. Something will be missing from many orders.
• Orders are scheduled. Most customers book a delivery window. Capacity per window depends on pickers and vehicles.
• Temperature matters. Frozen, chilled and ambient goods need different handling from shelf to doorstep.

Start with your fulfillment model

The biggest decision is who picks the order and where. Here are the four common models, from Instacart to Amazon Fresh and your own chain:

Regional chains usually start with in-store picking and add a dark store when one area's volume justifies it. Startups choose between a marketplace (light on assets, heavy on partnerships) and dark stores (heavy on assets, full control).

How to build a grocery delivery app, step by step

Nine steps. Step 2 starts on day one and runs in parallel with design, because catalog and stock data set the pace.

1. Pick the model and a pilot storeOne store or one dark store, one delivery area, pickup and delivery or pickup only. A single site lets you fix picking and slots before you copy them.
2. Audit catalog and inventory dataExport the product list from your POS or ERP. Check names, sizes, images, allergens, weighted items and how often on-hand counts match the shelf. Plan the cleanup and the feed that keeps it current.
3. Write substitution and weight rulesWhich items can be swapped, for what and at what price difference. How much over or under the ordered weight is acceptable. These become settings in the picker app, not judgment calls.
4. Design slot capacityEach delivery window has a limit based on picker hours, vehicle stops and cold storage. Show only windows you can keep, and set a cutoff after which the order is locked for picking.
5. Design the customer app around listsGrocery shoppers repeat. Past orders, saved lists, quick reorder and fast search by brand matter more than a pretty homepage. Show estimated prices for weighted items and explain the final charge.
6. Build the picker appAn optimized route through the aisles, scan to confirm each item, weight capture for variable items, substitution suggestions and live approval from the customer. Cold items get picked last.
7. Integrate POS, loyalty and paymentsPrices and promotions from the POS, loyalty points on online orders, card pre-authorization with final capture after picking, and SNAP EBT if your customers use it.
8. Pilot with staff ordersRun real orders from employees for a few weeks. Measure fill rate (items delivered as ordered), substitution acceptance, picking time per item and on-time delivery.
9. Add stores one at a timeEach new store brings new stock accuracy and layout. Onboard them with a checklist and watch the same metrics.

What you build and what you rent

Grocers already own most of the data. The build connects it to apps and makes it accurate enough to sell online. A typical split:

For an online grocery marketplace in Canada and the US, where stores sell food that is about to expire at a discount, our team built native apps for store employees and for buyers plus the backend. The employee app scans barcodes through dedicated hardware and includes a mini admin for offers and inventory. The data model had to hold unusual attributes too, such as acidity for wine or softness for bread.

Key technical rules for grocery

Most grocery app complaints come down to three things: something was missing, the price was different, or the window was missed. These rules address all three:

• Model weighted items properly. Store a unit price per pound, an estimated weight per item and the actual weight after picking. Many weighted items carry price-embedded barcodes from the store scale, and the picker app should read them.
• Treat stock as a probability. Show items as available only above a safety buffer, and learn from picks: if pickers keep marking an item missing, hide it sooner.
• Substitutions as a workflow, not a note. Preferences per item, suggestions ranked by brand, size and price, live approval in the customer app, and a default action when the customer doesn't answer in time.
• Hold, then capture. Authorize the estimated total plus a buffer for weight and swaps, then capture the actual amount after picking. Show the customer why the final number changed.
• Slots from real capacity. Calculate slot limits from picker hours, items per order and vehicle stops. Lock orders for editing at a cutoff so pickers aren't chasing changes.
• Keep the cold chain visible. Tag items by temperature zone, pick frozen and chilled last, stage them in the right storage and record it. Drivers should see which bags go first.
• Same prices and offers as the store. Pull promotions and loyalty pricing from the POS, or decide openly that online prices differ. Silent mismatches destroy trust fast.

Built by Gilzor

Results we’ve shipped

70+products launched

98%delivered on time

85%clients come back

Retail · Startup700+stores aggregatedMobile application for Grocery shops→Food delivery3.5 moto MVP launchOptimizing and controlling delivery for a Pizzeria Chain→EdTech4 moto a full iOS app on the App StoreA mobile app to develop brain functions→

Talk to the people who build it. Tell us about your project and get a free estimate of scope, timeline and cost.

See how we’d approach yours →

What goes into the first version

A grocery MVP should do the weekly shop well in one store. A typical split:

On platforms: one cross-platform codebase works well for the customer app and usually for the picker app. Choose native for the picker app only if it runs on rugged scanners with vendor-specific barcode SDKs.

Timeline and budget at a glance

A single store or small chain costs about $50k–100k. Instacart-style marketplaces and dark-store quick commerce start around $220k, pass $500k and take 10–16 months, released in stages. US onshore teams charge roughly 2.5–3 times more. For the full breakdown and a calculator, see our grocery delivery app development cost guide.

Mistakes that cost the most later

• Trusting POS stock counts as they are. Selling what isn't on the shelf leads to missing items, refunds and angry reviews. Add buffers and learn from picks.
• Charging only the checkout total. Without a hold buffer and final capture, every heavier steak is a loss and every swap is a manual refund.
• Offering slots you can't keep. Unlimited windows look generous until Saturday morning. Cap them from real picker and van capacity.
• Leaving substitutions to the picker's judgment. Customers get oat milk instead of whole milk and never order again. Rules, suggestions and live approval fix most of it.
• Picking frozen goods first. A route that starts in the freezer aisle delivers soft ice cream. Route cold items last and stage them cold.
• Launching in many stores at once. Each store has its own layout and stock accuracy. Prove the playbook in one, then copy it.

Grocery app readiness checklist

Tick what is already true for your project. It shows how close you are to taking real orders from one store.

FAQ

Where Gilzor fits

We build mobile apps for retail and e-commerce, with the backend, integrations and QA behind them. For an online grocery marketplace in Canada and the US, our team built native iOS and Android apps for buyers and store employees: barcode scanning, offer and inventory management, store discovery on a map, checkout through Stripe, real-time order tracking, a loyalty system and three languages. We work from Poland and Cyprus, with a few shared hours a day with the US East Coast.

Send us your store count, your POS and how you want to pick and deliver orders. We'll help you choose the fulfillment model, plan the data work and cut a first version you can pilot in one store.